Case control studies Flashcards
What are GWAS
genome wide association studies
What is are Bradford - Hill criteria
minimum conditions necessary to provide adequate evidence of a causal relationship between incidence and outcome
1 strength of assoc 2 consistency (of study designs, populations) 3 specificity 4 temporality 5 dose response relationship 6 biological plausibility 7 coherence 8 experimental evidence 9 analogy (alternative explanations)
Why might there be inconsistencies between reports of associations between genotype and outcome?
Variation of underlying associations in populations (different ethnicities have different variants of genes which can effect outcome)
Heterogenous phenotype
confounding factors
chance
publication bias
What are the advantages/disadvantages of a case control study
inexpensive and quick
good for quick outcomes
not good for rare exposures
prone to selection recall bias
no estimates for disease incidence
Why should controls in a study come from the same population as the cases?
Same confounding factors; reduces differences in risk factors between groups
How do you calculate odds ratio
Ratio of disease odds in exposed group to disease odds in unexposed group
say for smokers and cancer
(smokers who develop cancer / smokers who don’t) / (non smokers who develop cancer / non smokers who don’t)
Shows that someone who has cancer is 2.25x more likely to be a smoker
Why is selection bias of controls important
Different groups have different incidence rates for diseases.
Say if you were going to use a hospital ward as the control group for a trial - the ward you pick will have different incidences of diseases from a different ward (e.g trauma ward vs non-trauma ward)
So this affects odds ratio in final result
What is diagnostic bias and what effect can it have
where a treatment process leads to an increased detection of the disease (like people on birthcontrol dont get periods, so when they bleed it is a sign of cancer - but if she wasnt on the birthcontrol the bleeding wouldnt necessarily be considered abnormal so the cancer wouldn’t be detected) so this leads to an INCREASE IN OR
What is survival bias
Only patients who have survived a lethal disease can enter the study DECREASE IN OR
What is admission bias
exposed cases more likely to be admitted than controls INCREASE IN OR
What is non-response bias
people not taking treatment/test (e.g not turning up to a scan) DECREASE IN OR
What is recall bias
people in exposed group remember events better than controls INCREASE IN OR
What is interviewer bias
different questions asked to either group
